1S77.J OlO [KOnig. 



I called attention to the effect of vanadium on the titanium reaction 

 some time ago (Proceedings Acad. Nat. Sciences, Marcli, 1876), when it 

 was stated that the two metals, producing complimentary colors —red and 

 green — will extinguish each other ; vanadium, however, may be recognized 

 by its deep yellow color in the oxydizing flame, which is most intense at 

 low temperature, while the yellow of titanium only shows at high tempera- 

 ture, just below red heat. This point is not only of general interest, as a 

 case which might possibly present itself to the analyst, but is of direct 

 practical importance, since I have found vanadium in all the titanium 

 minerals of Magnet Cove (Proceedings Acad. Nat. Sciences, March, 1876), 

 and became aware of its existence just by applying this colorimetric 

 method, to the small residues obtained from the iron sesquioxyd, by 

 means of citric acid and ammonium sulphide. Here the titanium had 

 escaped precipitation by boiling, and would have been called alumina, 

 since it formed a nearly white powder and did not give the amethyst red 

 color in the reducing flame. However, vanadium and chromium may be 

 removed from the oxyds by fusion with potassium or sodium nitrate, a 

 treatment -which should never be omitted, if the highest degree of accuracy 

 is desired in analysis. 



Practical execution of the metJiod. I prepare titanium dioxyd and 

 aluminum sesquioxyd, the former from rutile, the latter from alum, in the 

 finest possible division, by igniting their respective ammonium citrates or 

 tartrates, removing at the same time the last trace of iron by ammonium 

 sulphide. In this fine condition the oxyds are very readily dissolved by 

 the acid sodium phosphate before the blow-pipe. I tlien weigh out 



0.1 gr. ; 0.095 gr. ; 0.090 gr. ; 0.085 gr 0.005 gr. of each of the oxyds, 



and mix them together inversely by very careful trituration in the mortar. 

 I produce thus one series containing in 



TiO, AlA 



100 



95 5 



90 10 



85 15 



100 



These twenty samples, of one decigramme each, I put in well corked 

 glass tubes with their respective number ; they form the standard. 0.003 gr. 

 of the standard mixture, and an equal weight of the unknown mixture are 

 dissolved in equal quantities of microcosmic salt. For very accurate de- 

 termination it is necessary to weigh the salt (0.12 gr.), but with some 

 experience it is sufficient to form a circular loop on the platinum wire one- 

 eighth of an inch in diameter, then to add gradually of the salt until the 

 bead is nearly spherical, possessing one-eighth inch thickness. I use a 

 guage for this purpose formed of two pins, whose points are one-eighth 



